Guitar-Chords-recognition
A ML Project to classify guitar chords using CNN.
Dataset
The chords dataset was collected from MONTEFIORE RESEARCH GROUP of University of Liège - Montefiore Institute (Montefiore.ulg.ac.be, 2019). The chords dataset consists of 10 types of chords with 200 audio files of each chord.
download link: http://www.montefiore.ulg.ac.be/services/acous/STSI/file/jim2012Chords.zip
Libraries required:
- numpy
- keras
- librosa
- tensorflow
- pandas
- tkinter
- pygame
How to run:
Just install the required libraries and run classify.py
chord_recognition.ipynb
It creates the trained model model.json for prediction of guitar chords
classify.py
It uses the trained model model.json to predict a recorded guitar chord using a simple UI. In the UI, just press reocrd and play a chord. It records for 3 seconds and saves the output wav file recoeded.wav. The classify button shows the predicted chord.
